Semantic clustering technique in South Asian dictionary

[SILK at AC.GRIN.EDU]

While loath to get involved in this I fear largely fruitless conversation,
allow me to ask one question.  Dr. Kalyanaraman writes:

>     I believe, that it is not necessary to establish 'ancestry' for a word.
>If
>     it is found across scores of languages spread across vast distances, and
>
>     authenticated in very, very ancient literary texts and epigraphs, it does
>
>     not really matter which phonetic variant came first, despite Mayrhoffer
>and
>     Burrow/Emeneau disagreeing. What is more important are the 'images'
>
>     associated with or evoked by the phonemic variants of a language-family.
>
>

What exactly is meant here by "important"?  We need not launch into a long
discussion of etymology vs. usage, or invoke Bhart.rhari and so forth, I
think, if we clarify first and foremost what it is we are after.  What does
each participant in this discussion mean by "important"?  I suspect we will
discover that in the different answers to that question lie the fundamental
disagreements we have seen recently.

Jonathan Silk
SILK at AC.GRIN.EDU